English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Federated Learning (FL) enables training a global model without sharing thedecentralized raw data stored on multiple devices to protect data privacy. Dueto the diverse capacity of the devices, FL frameworks struggle to tackle theproblems of straggler effects and outdated models. In addition, the dataheterogeneity incurs severe accuracy degradation of the global model in the FLtraining process. To address aforementioned issues, we propose a hierarchicalsynchronous FL framework, i.e., FedHiSyn. FedHiSyn first clusters all availabledevices into a small number of categories based on their computing capacity.After a certain interval of local training, the models trained in differentcategories are simultaneously uploaded to a central server. Within a singlecategory, the devices communicate the local updated model weights to each otherbased on a ring topology. As the efficiency of training in the ring topologyprefers devices with homogeneous resources, the classification based on thecomputing capacity mitigates the impact of straggler effects. Besides, thecombination of the synchronous update of multiple categories and the devicecommunication within a single category help address the data heterogeneityissue while achieving high accuracy. We evaluate the proposed framework basedon MNIST, EMNIST, CIFAR10 and CIFAR100 datasets and diverse heterogeneoussettings of devices. Experimental results show that FedHiSyn outperforms sixbaseline methods, e.g., FedAvg, SCAFFOLD, and FedAT, in terms of trainingaccuracy and efficiency.

FedHiSyn: A Hierarchical Synchronous Federated Learning Framework for Resource and Data Heterogeneity